
import pandas as pd
import networkx as nx
import re
import yaml
from itertools import chain
from pathlib import Path
from operator import itemgetter
from collections import defaultdict

import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np

#Import data_tools module 
import sys
sys.path.insert(0, '/home/agonzalez/data_tools/')

from data_tools.files import download
#from path_plots.dmdb_plots import plot_path
from data_tools.plotting import count_plot_h, venn2_pretty, venn3_pretty

from pySankey.sankey import sankey


# Make the output folders
this_name = '1_basic_dmdb_analysis'
out_dir = Path('../2_pipeline').joinpath(this_name, 'out').resolve()
out_dir.mkdir(parents=True, exist_ok=True)

data_dir = Path('../0_data/external').resolve()
data_dir.mkdir(parents=True, exist_ok=True)


#DONT RUN THIS AGAIN. LAST TIME IT WAS SAVED: APRIL 11 2022
#DMDB_URL = 'https://raw.githubusercontent.com/SuLab/DrugMechDB/main/indication_paths.yaml'
#download(DMDB_URL, data_dir.joinpath('indication_paths.yaml'), redownload=False)


with open(data_dir.joinpath('indication_paths.yaml'), 'r') as fh:
        ind = yaml.safe_load(fh)


def path_to_tup(path):
    return (path['graph']['drugbank'], path['graph']['disease_mesh'])

def path_to_G(path):
    return nx.node_link_graph(path)                                                                                        

def get_all_paths(path):
    source_id = path['links'][0]['source']                                                                              
    target_ids = list(set([l['target'] for l in path['links']]) - set([l['source'] for l in path['links']]))
    G = path_to_G(path)
    this_paths = list(chain(*[list(nx.all_simple_paths(G, source_id, target_id)) for target_id in target_ids]))         
    return this_paths

def get_id_to_type(G):
    id_to_type = {}
    for n in G.nodes.data():
        id_to_type[n[0]] = n[1]['label']
    return id_to_type

def get_id_to_name(G):
    id_to_name = {}
    for n in G.nodes.data():
        id_to_name[n[0]] = n[1]['name']
    return id_to_name

def add_metaedges(G):
    id_to_type = get_id_to_type(G)
    for e in G.edges:
        G.edges[e]['metaedge'] = id_to_type[e[0]] + ' - ' + e[2] + ' - ' + id_to_type[e[1]]
    return G

def add_meanode_pairs(G):
    id_to_type = get_id_to_type(G)
    for e in G.edges:
        G.edges[e]['mn_pair'] = id_to_type[e[0]] + ' - ' + id_to_type[e[1]]
    return G

def get_targets(G):
    drug = list(G.edges)[0][0]
    targets = []
    for e in G.edges:
        if e[0] == drug:
            targets.append(e[1])
    return targets

def get_target_metaedges(G):
    drug = list(G.edges)[0][0]
    target_mes = []
    if 'metaedge' not in G.edges[list(G.edges)[0]]:
        G = add_metaedges(G)
    
    for e in G.edges:
        if e[0] == drug:
            target_mes.append(G.edges[e]['metaedge'])
    return target_mes


basic_stats1 = defaultdict(list)
all_metaedges = []
all_parings = []
all_targets = []
unique_metaedges = []
first_edge_type = []
all_nodes = []

id_to_name = {}
id_to_label = {}

for i, p in enumerate(ind):
    drug_id, dis_id = path_to_tup(p)
    paths = get_all_paths(p)
    G = path_to_G(p)
    
    G = add_metaedges(G)
    G = add_meanode_pairs(G)
    
    basic_stats1['idx'].append(i)
    basic_stats1['id'].append(p['graph']['_id'])
    basic_stats1['drug'].append(drug_id)
    basic_stats1['disease'].append(dis_id)
    basic_stats1['n_nodes'].append(len(G.nodes))
    basic_stats1['n_edges'].append(len(G.edges))
    basic_stats1['n_paths'].append(len(paths))
    basic_stats1['longest_path'].append(max([len(p) for p in paths]))
    basic_stats1['shortest_path'].append(max([len(p) for p in paths]))
    basic_stats1['metapath'].append(" - ".join([n[1]['label'] for n in G.nodes.data()]))
    basic_stats1['metapath_with_edges'].append("".join([re.sub(" - [^-]*$"," - ",e[2]['metaedge']) for e in G.edges.data()])+"Disease")

    
    this_metaedges = [G.edges[e]['metaedge'] for e in G.edges]
    
    all_metaedges += this_metaedges
    unique_metaedges += list(set(this_metaedges))
    
    all_parings += [G.edges[e]['mn_pair'] for e in G.edges]
    all_targets += get_targets(G)
    first_edge_type += get_target_metaedges(G)
    all_nodes += list(G.nodes)
    
    id_to_label = {**id_to_label, **get_id_to_type(G)}
    id_to_name = {**id_to_name, **get_id_to_name(G)}
    
basic_stats1 = pd.DataFrame(basic_stats1)


basic_stats1.head()


#lets take an example
basic_stats1[basic_stats1["id"]=="DB00316_MESH_D010146_1"]["metapath"].iloc[0]

'Drug - Protein - Protein - Pathway - ChemicalSubstance - Disease'


basic_stats1[basic_stats1["id"]=="DB00316_MESH_D010146_1"]["metapath_with_edges"].iloc[0]

'Drug - decreases activity of - Drug - decreases activity of - Drug - negatively regulates - Protein - increases abundance of - Protein - increases abundance of - Pathway - affects risk for - ChemicalSubstance - causes - Disease'


all_metapath_nodes = defaultdict(list)
for i, p in enumerate(ind):
    
    _id = (p["graph"]["_id"])
    
    drug_id, dis_id = path_to_tup(p)
    paths = get_all_paths(p)
    G = path_to_G(p)
    
    G = add_metaedges(G)
    G = add_meanode_pairs(G)
    
    this_metaedges = [G.edges[e]['metaedge'] for e in G.edges]
   
   
    
    #lets construct a graph for metapath
    graph = nx.DiGraph()
    for rel in this_metaedges:
        rel = rel.split(" - ")
        graph.add_edge(rel[0], rel[2], weight =rel[1]) #Add nodes and edges 
        
    edge_labels = nx.get_edge_attributes(graph,'weight') #set edge labels
    
    
    #get paths (nodes) and #add the edges     
   
    
    try: 
        st = 'Drug'  #start node
        end = 'Disease'  #end node
        for mp in nx.all_simple_paths(graph, st, end):
            mp_str = (" - ".join(mp))
            
            all_metapath_nodes[_id].append(mp_str)
            
        
    except:
        st = 'Disease'  #start node
        end = 'Drug'  #end node
        for mp in nx.all_simple_paths(graph, st, end):
            mp_str = (" - ".join(mp))
            all_metapath_nodes[_id].append(mp_str)


all_metapath_edges = defaultdict(list)
for i, p in enumerate(ind):
    _id = (p["graph"]["_id"])
    drug_id, dis_id = path_to_tup(p)
    paths = get_all_paths(p)
    G = path_to_G(p)
    
    G = add_metaedges(G)
    G = add_meanode_pairs(G)
    
    this_metaedges = [G.edges[e]['metaedge'] for e in G.edges]
   
   
    
    #lets construct a graph for metapath
    graph = nx.DiGraph()
    for rel in this_metaedges:
        rel = rel.split(" - ")
        graph.add_edge(rel[0], rel[2], weight =rel[1]) #Add nodes and edges 
        
    edge_labels = nx.get_edge_attributes(graph,'weight') #set edge labels
    
    
    #get paths (nodes) and #add the edges     
    try:
        st = 'Drug'  #start node
        end = 'Disease'  #end node
        meta_path=[]
        for mp in nx.all_simple_paths(graph, st, end):
            
            metaedge =[]

            #for value in mp 
            for n in range(len(mp)):
                if n == (len(mp)-1):
                    continue

                else:
                     result = (mp[n] + " - " + edge_labels.get((mp[n], (mp[n+1])))+ " - ")
                     metaedge.append(result)
            
            me = ("".join(metaedge)+ "Disease")
            all_metapath_edges[_id].append(me)

    except: 
        st = 'Disease'  #start node
        end = 'Drug'  #end node
        meta_path=[]
        for mp in nx.all_simple_paths(graph, st, end):
            
            metaedge =[]

            #for value in mp 
            for n in range(len(mp)):
                if n == (len(mp)-1):
                    continue

                else:
                     result = (mp[n] + " - " + edge_labels.get((mp[n], (mp[n+1])))+ " - ")
                     metaedge.append(result)
            
            me = ("".join(metaedge)+ "Drug")
            all_metapath_edges[_id].append(me)


basic_stats = defaultdict(list)
all_metaedges = []
all_parings = []
all_targets = []
unique_metaedges = []
first_edge_type = []
all_nodes = []

id_to_name = {}
id_to_label = {}

for i, p in enumerate(ind):
    _id = (p["graph"]["_id"])
    drug_id, dis_id = path_to_tup(p)
    paths = get_all_paths(p)
    G = path_to_G(p)
    
    G = add_metaedges(G)
    G = add_meanode_pairs(G)
    
    basic_stats['idx'].append(i)
    basic_stats['id'].append(p['graph']['_id'])
    basic_stats['drug'].append(drug_id)
    basic_stats['disease'].append(dis_id)
    basic_stats['n_nodes'].append(len(G.nodes))
    basic_stats['n_edges'].append(len(G.edges))
    basic_stats['n_paths'].append(len(paths))
    basic_stats['longest_path'].append(max([len(p) for p in paths]))
    basic_stats['shortest_path'].append(max([len(p) for p in paths]))
    basic_stats['metapath'].append(all_metapath_nodes[_id])
    basic_stats['metapath_with_edges'].append(all_metapath_edges[_id])

    
    this_metaedges = [G.edges[e]['metaedge'] for e in G.edges]
    
    all_metaedges += this_metaedges
    unique_metaedges += list(set(this_metaedges))
    
    all_parings += [G.edges[e]['mn_pair'] for e in G.edges]
    all_targets += get_targets(G)
    first_edge_type += get_target_metaedges(G)
    all_nodes += list(G.nodes)
    
    id_to_label = {**id_to_label, **get_id_to_type(G)}
    id_to_name = {**id_to_name, **get_id_to_name(G)}
    
basic_stats = pd.DataFrame(basic_stats)


basic_stats


print("Number of unique drugs: " + str(len(set(basic_stats['drug']))))
print("Number of unique diseases: " + str(len(set(basic_stats['disease']))))
print("Number of unique drug-disease combinations: " + str(len(basic_stats.groupby(['drug','disease']).size())))

Number of unique drugs: 1525
Number of unique diseases: 677
Number of unique drug-disease combinations: 3467


sns.histplot(basic_stats['n_nodes'],bins=20)
plt.xlim(2, 20)
plt.title('Distribution of nodes per indications')
plt.xlabel('Number of Nodes')
plt.ylabel('Number of indications');


sns.histplot(basic_stats['n_edges'],binwidth=1)
plt.title('Distribution of edges per indication')
plt.xlabel('Number of Edges')
plt.ylabel('Number of indications');


max(basic_stats['n_paths'])

40


g = sns.histplot(basic_stats['n_paths'],bins=10)
g.set_yscale('log')
plt.title('Distribution of paths per indication')
plt.xlabel('Number of Paths')
plt.ylabel('Number of indications');


sns.histplot(basic_stats['longest_path'],binwidth=1)
plt.title('Distribution of longest path per record')
plt.xlabel('Length of longest path')
plt.ylabel('Number of Records');


pair_counts = pd.Series(all_parings).value_counts()
pair_counts

Drug - Protein                           3650
Protein - BiologicalProcess              2548
BiologicalProcess - Disease              1862
BiologicalProcess - BiologicalProcess    1805
PhenotypicFeature - Disease               960
                                         ... 
Cell - OrganismTaxon                        1
Disease - GeneFamily                        1
PhenotypicFeature - Protein                 1
OrganismTaxon - ChemicalSubstance           1
MacromolecularComplex - GeneFamily          1
Length: 140, dtype: int64


pair_counts = pd.Series(all_parings).value_counts()
count_plot_h(pair_counts.head(10))
plt.title('Most common Metanode Pairings in DrugMechDB')
plt.xlabel('Number of times metanode pairing appears');


pd.Series(all_parings).value_counts()

Drug - Protein                           3650
Protein - BiologicalProcess              2548
BiologicalProcess - Disease              1862
BiologicalProcess - BiologicalProcess    1805
PhenotypicFeature - Disease               960
                                         ... 
Cell - OrganismTaxon                        1
Disease - GeneFamily                        1
PhenotypicFeature - Protein                 1
OrganismTaxon - ChemicalSubstance           1
MacromolecularComplex - GeneFamily          1
Length: 140, dtype: int64


#Create df for pairs and count them 
all_parings_set = set(all_parings)

pairs_df = pd.DataFrame()
num_pairs_nodes =[]
source_node = []
target_node = []

for pair in all_parings_set:
    num_pairs = len([i for i in all_parings if i == pair])
    num_pairs_nodes.append(num_pairs)
    
    value = pair.split(" - ")
    source_node.append(value[0])
    target_node.append(value[1])
        
pairs_df["source_node"]=source_node
pairs_df["target_node"]=target_node
pairs_df["count"]=num_pairs_nodes


pairs_df_hm =pairs_df.pivot("source_node", "target_node", "count")
pairs_df_hm= pairs_df_hm.fillna(0)


pairs_df_hm


sns.set(rc={"figure.dpi":100})
ax = sns.heatmap(pairs_df_hm, cmap= "viridis", vmin = 0, vmax = 1000)


#Same heatmap using log scale
sns.set(rc={"figure.dpi":100})
ax = sns.heatmap(np.log10(pairs_df_hm+1e-10), cmap= "viridis", vmin = 0, vmax = 3)


source =[]
target =[]

for val in all_parings:
    val = val.split(" - ")
    source.append(val[0])
    target.append(val[1])


sankey(source, target, aspect=15, fontsize=7)
# Get current figure
fig = plt.gcf()

# Set size in inches
fig.set_size_inches(6, 6)

# Set the color of the background to white
fig.set_facecolor("w")

# Save the figure
#fig.savefig("DMDB_relations_sankey.png", bbox_inches="tight", dpi=500)


me_counts = pd.Series(all_metaedges).value_counts()
count_plot_h(me_counts.head(10))
plt.title('Most common Metaedge in DrugMechDB')
plt.xlabel('Number of times Metaedge appears');


#Unique metaedges
ume_counts = pd.Series(unique_metaedges).value_counts()
count_plot_h(ume_counts.head(10))
plt.title('Most common Metaedge in DrugMechDB')
plt.xlabel('Number of records containing Metaedge');


len(set(unique_metaedges))

870


mp_nodes = []
for id_ in all_metapath_nodes:
    for p in all_metapath_nodes[id_]:
        mp_nodes.append(p)
freq_mp = dict((x,mp_nodes.count(x)) for x in set(mp_nodes))
freq_mp_sort = sorted(freq_mp.items(), key=lambda x: x[1],reverse = True)    

mp_nodes_df = pd.DataFrame()
mp_list = []
mp_count = []

for mp in freq_mp_sort:
    mp_list.append(mp[0])
    mp_count.append(mp[1])

mp_nodes_df["meta_path"] = mp_list
mp_nodes_df["count"] = mp_count


sns.set(rc={"figure.dpi":300})
sns.set_style("ticks") 
g = sns.barplot(x="count", y="meta_path", data=mp_nodes_df.head(10), orient = "h")
g.set_xlim(0,700)

for container in g.containers:
    g.bar_label(container,size= 10)


#Number of metapaths
len(set(mp_nodes))

313


mp_edges = []
for id_ in all_metapath_edges:
    for p in all_metapath_edges[id_]:
        mp_edges.append(p)
freq_mp = dict((x,mp_edges.count(x)) for x in set(mp_edges))
freq_mp_sort = sorted(freq_mp.items(), key=lambda x: x[1],reverse = True)    


mp_nodes_edges_df = pd.DataFrame()
mp_list = []
mp_count = []

for mp in freq_mp_sort:
    mp_list.append(mp[0])
    mp_count.append(mp[1])

mp_nodes_edges_df["meta_path-edges"] = mp_list
mp_nodes_edges_df["count"] = mp_count


sns.set(rc={"figure.dpi":300})
sns.barplot(x="count", y="meta_path-edges", data=mp_nodes_edges_df.head(10), orient = "h")
sns.set_style("white")


#number of meta_paths (with edges)
len(set(mp_edges))

1500


node_counts = pd.Series(all_nodes, name='id')
node_counts = node_counts.to_frame().reset_index()
node_counts['name'] = node_counts['id'].map(id_to_name)
node_counts['label'] = node_counts['id'].map(id_to_label)


count_plot_h(node_counts['name'].value_counts().head(10))
plt.title('Top 10 most encountered nodes')
plt.xlabel('Number of paths containing Node')
plt.ylabel('');


count_plot_h(node_counts['label'].value_counts().head(10))
plt.title('Top 10 most encountered node types')
plt.xlabel('Number of times node type encountered')
plt.ylabel('');


target_counts = pd.Series(all_targets, name='count')
target_counts = target_counts.value_counts().to_frame().reset_index().rename(columns={'index': 'id'})
target_counts['name'] = target_counts['id'].map(id_to_name)
target_counts['label'] = target_counts['id'].map(id_to_label)


count_plot_h(target_counts.set_index('name').head(10)['count'])
plt.title('Top 10 most encountered targets')
plt.xlabel('Number of paths containing Target')
plt.ylabel('');


pd.Series(first_edge_type, name='type').to_frame()


fet_data = pd.Series(first_edge_type, name='type').to_frame()
fet_data['target_type'] = fet_data['type'].apply(lambda x: x.split(' - ')[2])
fet_data['interaction'] = fet_data['type'].apply(lambda x: x.split(' - ')[1])


fet_data.head()


#Plot the top 10 most encountered edge types
count_plot_h(fet_data['type'].value_counts().head(10))
plt.title('Top 10 most encountered targets')
plt.xlabel('Number of paths containing Target')
plt.ylabel('');


#Plot the top target label interactions

count_plot_h(fet_data['target_type'].value_counts())
plt.title('Types of targets in Drug-Target interactions')
plt.xlabel('Number Drug-Target interactions')
plt.ylabel('');


#Lets plot the interaction types connecting Drug-Target
count_plot_h(fet_data['interaction'].value_counts().head(10))
plt.title('Top 10 Drug to Target interaction semantics')
plt.xlabel('Number of Drug-Target interactions')
plt.ylabel('');


for p in ind:
    G = path_to_G(p)


#Build graph 
Graph = nx.MultiDiGraph()

for p in ind:
    G = path_to_G(p)
    Graph = nx.compose(Graph, G)


print(nx.info(Graph))

Name: 
Type: MultiDiGraph
Number of nodes: 4947
Number of edges: 10342
Average in degree:   2.0906
Average out degree:   2.0906


nx.density(Graph)

0.00042267689755647633


sgraph = nx.MultiGraph(Graph)


components = list(nx.connected_components(sgraph))
len(components)

20


larget_component = max(components, key=len)


subgraph = sgraph.subgraph(larget_component)


nx.diameter(subgraph)

25


degree_dict = dict(Graph.degree(Graph.nodes))
nx.set_node_attributes(Graph, degree_dict, 'degree')
sorted_degree = sorted(degree_dict.items(), key=itemgetter(1), reverse=True)


node_degree = pd.DataFrame(sorted_degree).rename(columns={0: 'id', 1: 'degree'})
node_degree['name'] = node_degree['id'].map(id_to_name)


sns.histplot(node_degree['degree'],binwidth=1,binrange=[0,50])
plt.title('Distribution of node degree')
plt.xlabel('Node degree')
plt.ylabel('Number of nodes');


count_plot_h(node_degree.set_index('name')['degree'].head(10))
plt.title('Degree of most connected nodes');
plt.xlabel('Degree')
plt.ylabel('');

UsageError: Cell magic `%%jupyter` not found.

	idx	id	drug	disease	n_nodes	n_edges	n_paths	longest_path	shortest_path	metapath	metapath_with_edges
0	0	DB00619_MESH_D015464_1	DB:DB00619	MESH:D015464	3	2	1	3	3	[Drug - Protein - Disease]	[Drug - decreases activity of - Protein - caus...
1	1	DB00619_MESH_D034721_1	DB:DB00619	MESH:D034721	5	5	2	4	4	[Drug - Protein - BiologicalProcess - Disease]	[Drug - decreases activity of - Protein - posi...
2	2	DB00316_MESH_D010146_1	DB:DB00316	MESH:D010146	6	7	3	4	4	[Drug - Protein - ChemicalSubstance - Disease,...	[Drug - decreases activity of - Protein - incr...
3	3	DB00316_MESH_D005334_1	DB:DB00316	MESH:D005334	5	4	1	5	5	[Drug - Pathway - GrossAnatomicalStructure - B...	[Drug - negatively regulates - Pathway - occur...
4	4	DB00945_MESH_D010146_1	DB:DB00945	MESH:D010146	5	5	2	4	4	[Drug - Protein - ChemicalSubstance - Disease]	[Drug - decreases activity of - Protein - incr...
...	...	...	...	...	...	...	...	...	...	...	...
3733	3733	DB00618_MESH_D013203_1	DB:DB00618	MESH:D013203	7	7	2	6	6	[Drug - CellularComponent - BiologicalProcess ...	[Drug - decreases activity of - CellularCompon...
3734	3734	DB00618_MESH_D013587_1	DB:DB00618	MESH:D013587	7	7	2	6	6	[Drug - CellularComponent - BiologicalProcess ...	[Drug - decreases activity of - CellularCompon...
3735	3735	DB00618_MESH_D014141_1	DB:DB00618	MESH:D014141	7	7	2	6	6	[Drug - CellularComponent - BiologicalProcess ...	[Drug - decreases activity of - CellularCompon...
3736	3736	DB00618_MESH_D014552_1	DB:DB00618	MESH:D014552	7	7	2	6	6	[Drug - CellularComponent - BiologicalProcess ...	[Drug - decreases activity of - CellularCompon...
3737	3737	DB00618_MESH_D015001_1	DB:DB00618	MESH:D015001	7	7	2	6	6	[Drug - CellularComponent - BiologicalProcess ...	[Drug - decreases activity of - CellularCompon...

target_node	BiologicalProcess	Cell	CellularComponent	ChemicalSubstance	Disease	Drug	GeneFamily	GrossAnatomicalStructure	MacromolecularComplex	MolecularActivity	OrganismTaxon	Pathway	PhenotypicFeature	Protein
source_node
BiologicalProcess	1805.0	158.0	111.0	485.0	1862.0	17.0	8.0	203.0	2.0	6.0	666.0	26.0	719.0	48.0
Cell	74.0	0.0	0.0	5.0	110.0	2.0	0.0	4.0	0.0	0.0	1.0	0.0	5.0	3.0
CellularComponent	105.0	3.0	48.0	4.0	18.0	0.0	2.0	6.0	0.0	17.0	138.0	0.0	10.0	16.0
ChemicalSubstance	871.0	19.0	90.0	82.0	407.0	6.0	80.0	79.0	3.0	11.0	39.0	7.0	169.0	258.0
Disease	7.0	0.0	0.0	2.0	111.0	0.0	1.0	3.0	0.0	0.0	0.0	0.0	5.0	14.0
Drug	349.0	8.0	87.0	628.0	70.0	46.0	560.0	23.0	13.0	53.0	23.0	19.0	18.0	3650.0
GeneFamily	527.0	3.0	7.0	71.0	35.0	0.0	13.0	5.0	0.0	45.0	4.0	19.0	15.0	32.0
GrossAnatomicalStructure	111.0	0.0	3.0	8.0	173.0	0.0	4.0	58.0	0.0	3.0	0.0	2.0	91.0	4.0
MacromolecularComplex	18.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
MolecularActivity	316.0	12.0	3.0	35.0	7.0	0.0	3.0	4.0	0.0	8.0	31.0	4.0	6.0	4.0
OrganismTaxon	10.0	0.0	0.0	1.0	818.0	2.0	0.0	0.0	0.0	0.0	0.0	0.0	1.0	0.0
Pathway	219.0	2.0	0.0	99.0	15.0	0.0	1.0	2.0	0.0	0.0	4.0	0.0	13.0	19.0
PhenotypicFeature	20.0	1.0	0.0	2.0	960.0	7.0	0.0	14.0	0.0	0.0	10.0	0.0	48.0	1.0
Protein	2548.0	39.0	53.0	739.0	48.0	7.0	72.0	59.0	0.0	386.0	6.0	388.0	81.0	488.0

	type
0	Drug - decreases activity of - Protein
1	Drug - decreases activity of - Protein
2	Drug - decreases activity of - Protein
3	Drug - decreases activity of - Protein
4	Drug - decreases activity of - Protein
...	...
5481	Drug - decreases activity of - CellularComponent
5482	Drug - decreases activity of - CellularComponent
5483	Drug - decreases activity of - CellularComponent
5484	Drug - decreases activity of - CellularComponent
5485	Drug - decreases activity of - CellularComponent

Basic DMDB Analysis¶

Things we'll explore in this notebook¶

Metanode parings¶

Metaedge use¶

Metaedges per record¶

Metapaths¶

Node information¶

Target Information¶

Full Graph Analysis¶

	type	target_type	interaction
0	Drug - decreases activity of - Protein	Protein	decreases activity of
1	Drug - decreases activity of - Protein	Protein	decreases activity of
2	Drug - decreases activity of - Protein	Protein	decreases activity of
3	Drug - decreases activity of - Protein	Protein	decreases activity of
4	Drug - decreases activity of - Protein	Protein	decreases activity of